Anthropic Changes Messages API to Protect Against Distillation
maksym_andr · x · 2026-09-02
To prevent industrial-scale illicit distillation, Anthropic has changed how the Messages API handles 'thinking blocks' in Claude Fable 5.1. New API accounts can no longer modify the context (messages, tools, or system prompt) preceding a thinking block during multi-turn conversations.
The API now verifies that a thinking block is returned with the exact same system prompt, tools, and messages that produced it, returning an error if they don't match. Developers can opt-in to have thinking blocks removed from such requests. This currently applies to new API accounts only but will extend to all in future releases.
More from Infra
- Energy Throttling LLMs via MLP Contextual Bandit — hongyangzh · 2026-09-02
- PgDog vs. RDS Proxy: Open Source Alternative is 2x Faster — _changxu · 2026-09-02
- NVIDIA revenue up 1,730% in 4 years, targeting $637B next year — thedealdirector · 2026-09-02
- Deploying DeepSeek-V4 on Blackwell: Fixing 3 critical SGLang bugs — shrug_hellifino · 2026-09-02
- LangChain fine-tunes Qwen for agent evals, cutting costs by 100x vs GPT-5.5 — LangChain · 2026-09-02
- B300 Supports 6x More Concurrent Agents Than H200: Benchmark — ryanshrout · 2026-09-02